Contents
Why this matters more than the spread
Spread is visible, it is on the screen, and it is obviously a cost. Swap is invisible, it is charged overnight without a prompt, and for most retail traders it is the largest single cost in the system. A trader holding gold for five nights at 1 lot on a standard account can pay more in swap than they paid in spread across the same trade.
The direction matters too. On most retail accounts, long gold positions pay negative swap and short positions receive positive swap, because the broker implicitly charges for holding a long non-yielding position and pays for the convenience of holding a short. This means a carry cost that is structurally against long positions.
Our swap calculator guide covers the mechanics, and carry trade analysis covers the strategy side.
How the numbers work
Swap is quoted in points per lot per night, and it is calculated on the notional value of the position, which for gold is the contract size times the price. For XAUUSD on a standard contract of 100 ounces, a price of 3,400 means a notional of 340,000, and a swap rate of -25 points per lot per night is $25 per lot per night on a roughly 340,000 notional.
The rate is not constant across brokers and it changes with the central bank policy that determines the underlying rate differential, which means it moves over weeks and months. A strategy that worked in a high-rate environment can become unprofitable in a low-rate one purely through this channel, with no change to the price behaviour of gold.
The Wednesday charge is also worth noting. Most brokers charge three days of swap on Wednesday night, covering the weekend. This is a real cost and it means a five-night hold is not five daily charges but six.
The strategies that fail because of it
The first is the naive swing. A trader takes a position on a daily or four-hour chart, holds it for a week, and pays a week of negative swap on a position whose expected move may only be a few hundred points. A 25-point-per-night charge is 125 points over a week, which against a 300-point target is over 40 percent of the gross edge.
The second is the carry trade done backwards. The genuine carry trade in gold involves a specific expectation of the rate environment, and most retail versions of it are simply an unhedged long position that happens to earn a small negative swap, which is not a carry trade.
The third is the breakout hold. A breakout trade that produces a good move but is held for several sessions because the trader is waiting for more can easily give back the whole gain to swap, particularly in a ranging market where the breakout does not extend.
The fix in all three cases is the same: know the cost per night of the position, and set a maximum holding time as part of the setup rather than deciding in the moment. A position with a 25-point-per-night carry should either have a target that comfortably exceeds a few days of that, or be designed to be closed within the session.
How to incorporate swap into your edge calculation
Every trade in your expectancy calculation needs a cost term that includes spread, commission, and swap for the expected holding period. The expected holding period is the part people get wrong. If your rule is “exit when the target is hit or when the thesis is invalidated”, then the expected holding time is the average of those two outcomes, and the swap cost follows from that average rather than from the worst case.
The practical method is to run your backtest with three separate cost assumptions: a base case using the current swap rate, a stressed case at twice that rate, and a low-rate-environment case at zero. If the edge survives all three, it is reasonably robust to the rate environment. If it only works at zero swap, you have a carry-dependent strategy and you should know that.
This connects to the broader point in spread and cost analysis and to the regime concern in regime detection. A change in the interest rate environment is a genuine regime change, and it can kill a strategy that is otherwise sound.
Reducing the cost where you can
The most effective lever is holding time, and the second is position size relative to the account. Beyond that, there are broker-level options worth checking. Some brokers offer reduced or zero swap on gold for certain account types or in the Islamic account structure, which is worth investigating if you hold positions overnight regularly, though the eligibility rules vary and the account has to genuinely qualify.
Another option is to align the holding period with the sessions where the edge actually appears. If your edge is intraday flow, then holding overnight for a wider target is trading a different strategy, and it should be justified as one rather than drifted into.
The last one is the simplest and the most neglected: check whether the swap rate has changed recently. Traders assume their costs are static, and when the rate environment shifts, the cost model they are working from is silently wrong.